1. Technical Field
This invention relates to an in vitro solid-phase competitive assay for detecting the presence of an antigenic analyte, or a binding fragment or precursor thereof in a biological sample utilizing a fusion protein of an antigenic peptide and a binding peptide, a solid support-bound antibody selectively binding the binding peptide, an antibody selectively binding the antigenic peptide, and optionally an antibody binding molecule. The present assay provides higher sensitivity and specificity than all other available assays.
2. Description of the Background
The determination of levels of different antigens in animal and human tissues took a definite turn with the development of immunoassays. The concept on which immunoassays are based is the quantitative binding of a known antigen in known quantities to an antibody in equally known quantities, and the binding of this antibody to the antigen to be used as a standard, and a comparison of this to a unknown sample comprising the antigen which will also be bound by the antibody. Antigen present in the unknown sample interfered with the initial binding of known quantities of antibody to known quantities of antigen. A key step of these assays is the separation of the bound form of the antibody or the antigen from its unbound form. Many configurations for this reaction have been proposed either as direct immunometric, competitive or displacement assays, and the like. However, to quantitate results it is in general needed to resort to hemagglutination assays, radioimmunoassays, enzyme-linked assays, and the like.
In general, in an immunoassay, a given analyte or antigen present in a animal or human tissue is or may be solubilized for mixing with the immunoassay system, and it is then compared to a solubilized known quantity of the analyte. The most common tissue analyte is blood, and more specifically serum and/or plasma from blood, but urine, cerebro-spinal fluid, different serum preparations and different animal and human tissues and fluids are also routinely assayed.
Some of the areas which have most benefited with the advent of immunoassays have been clinical chemistry, endocrinology and oncology. In endocrinology and clinical chemistry enzyme-linked assays and radioimmunoassays have been used to determine levels of hormones, proteins, and lipid metabolites, among other substances. In the field of oncology blood components, and some times tissue antigens or other molecules, indicate either the appearance of cancer or a pre-cancerous condition in animals or men. These molecules are routinely tested to monitor appearance, relapse, progression or regression of a cancer disease. These antigens or molecules are called cancer and/or tumor markers. For many years markers have been used for this purpose. An example thereof is the oncofetal antigen CEA which is used in the diagnosis of carcinomas, especially those of the colon. Other cancer markers include enzymes such as lactic dehydrogenase and alkaline phosphatase, metabolites such as prostaglandin and polyamine, proteins such as .alpha.-fetoprotein and human chorionic gonadotrophin, among others. Immunoassays of these cancer markers are now applied to the diagnosis and follow up of cancer patients.
These assays generally use as standard a partially or fully purified tissue antigen. In some occasions, however, polypeptides are synthesized in the laboratory for use as antigens. The more purified the antigenic substance used as standard for the immunoassay is, the more specific and trustworthy the assay.
A set of membrane-related antigens have been used for the diagnosis of breast cancer. The antigens were originally called human mammary epithelial antigens and antibodies to them were obtained by injection of human milk fat globule (HMFG) membranes to rabbits. These were polyconal antibodies called anti-human mammary epithelial (anti-HME) antibodies. The antibodies were prepared after repeated absorptions and were found to bind breast epithelial cells selectively. The discovery of this breast epithelial system of antigens opened many new immunologic opportunities in immunohistopathology, serum assays, radioimaging and eventually immunotherapy.
The anti-HME antibodies were shown to bind to breast epithelial cell lines as well as normal breast cells, but not to fibrocytes, vascular cells, and blood cells. A special group of breast epithelial antigens (BrE-antigens) originally called human mammary epithelial antigens (HME-ntigens), are bound by absorbed anti-serum (anti-HME serum) which were created in the rabbit. These antigens were found to have 150, 70 and 45-48 Kdalton molecular weights as established by affinity chromatography, and polyacrylamide gel electrophoresis double antibody immunoprecipitation. A similar system was shown to exist in the mouse. Mouse mammary epithelial antigens may also be detected by absorbed rabbit polyclonal antisera. These antisera also identify in the mouse mammary cell membrane components having molecular weights of 150, 70 and 45-48 Kdaltons. The antigens may be detected in either normal or neoplastic mouse mammary gland. These antigens are not detected in other normal tissue cells of mice.
Other polyclonal antisera were reported to have been produced against a step-purification of HMFG antigens. These antisera are pan-epithelial in nature and reactive only against the non-penetrating glycoprotein (NPGP) complex in contrast to the original anti-HME antibodies that bind the about 45, 70 and 150 kdalton antigens. Although the anti-HME antibodies bind before absorptions to the NPGP complex, anti-HME antibody final preparations do not recognize the NPGP complex as a result of absorptions with non-breast epithelial cells to render anti-HME specific.
HME antigens may be quantitated by an immunoassay in various human breast and non-breast cell lines and in normal breast epithelial cells. High concentrations of HME antigens were found in normal breast epithelial cells and in neoplastic cells. A protease treatment of live breast epithelial cell surfaces releases most antigens therefrom. Similar results show a 48-72 hour time lapse for full reconstitution of the normal breast epithelial cell membrane after digestion.
High levels of HME antigens are found in the sera of nude mice carrying human breast tumors. These antigens can be abolished by surgical removal of the breast tumor. Anti-HME antisera were shown to have certain specificity since other transplantable human tumors such as colon, lung and melanoma, did not increase HME antigen values in mice serum.
The specificity of the assay using anti-HME antigen serum for breast tumors was tested in a nude mouse model carrying transplantable human breast tumors and compared to the specificity of an assay for sialyl transferase levels, which is also a breast cancer marker. The levels of the enzyme which is present on the breast epithelial cell membrane and the HME antigens were measured simultaneously in the sera of nude mice grafted with human breast and non-breast tumors. Breast tumor-bearing mice had elevated levels of both serum markers. However, sialytransferase levels were also elevated in non-breast tumors while HME antigens were not. Upon surgical removal of all tumors, the presence of HME antigens declined precipitously in breast tumor-bearing nude mice while sialyl transferase levels remained elevated in both breast and non-breast tumor bearing animals. This is possibly due to surgical trauma and wound healing. The higher specificity of the HME antigen assay was thus proven at least in regard to sialyl transferase, a non-specific co-habitant of the cell membrane together with HME antigens. This indicates again that most, if not all, components of the breast epithelial cell are released into circulation by breast tumors, and that assay specificity, such as is obtained with an assay utilizing HME antigens, may be required to avoid that concurrent ailments or reactions in the tumor host interfere with the values obtained from sera with markers such as sialyl transferase.
HME antigens levels in the sera of breast cancer patients were also obtained using a slightly different radioimmunoassay (U.S. Pat. No. 4,584,268 to Ceriani and Peterson). In this assay, beads coated with polyclonal antibodies were incubated with a patient's serum, then the immobilized antigen was detected with the polyclonal antibodies labeled with biotin, and the latter detected by radiolabeled avidin. The assay was specific for positive cases of breast cancer since the sera of normal subjects, both male and female, suffering from benign diseases of the breast, carcinomas of lung and colon, neuroblastomas and melanomas yielded negative results. In contrast, 25% of Stage I primary breast carcinomas and more than 75% of disseminated breast cancer cases were found to have values above the cut off line.
To date the only complete proof of the existence of HME antigens, or any other breast epithelial antigens (BrE-antigens) in human sera with elevated values of the breast tumor markers, is provided by a very sensitive technique using in situ radioiodination of the HME antigens bound to an immobilized antibody. In contrast, only a small fraction of breast cancer patients, most of whom had elevated values of BrE antigens, gave positive results when less stringent criteria to detect BrE antigens in sera such as Western blotting were used. Elevated values of the three HMFG antigens detected by anti-HME antibodies were found in the circulation employing the in situ radioiodination approach. These antigens had 150, 70 and 45-48 Kdafton molecular weights in all breast cancer cases. Control sera from normal subjects and patients with colon and lung carcinomas were found to be negative. In addition, the antigen corresponding to one monoclonal antibody (Mc3) was also found in the sera of these patients by the in situ labeling technique. In later work, the Mc3 antigen was found to be associated with immune complexes in breast cancer patients.
Monoclonal antibodies have been used in immunoassays. However, their low binding constants and their restricted specificity are drawbacks to their use. Polyclonal antibodies, on the contrary, combine the specificities for several epitopes of the same antigen. Monoclonal antibodies were originally prepared against HMFG and also against breast tumor cells. As mentioned above, the most immunogenic of the HMFG antigens and breast cells is the NPGP complex, described for its binding to monoclonal antibody Mc1 (also called HMFG-2), and Mc5. This is the only antigen which has thus far been extensively quantitated in serum assays.
The original monoclonal antibodies against HMFG were followed by other monoclonal antibodies created in different laboratories. Immunoassays applied to obtain serum values for primary breast tumor patients and for disseminated disease patients yielded partially positive values in cases of primary breast tumors and small tumor loads. As the tumor load increased, more sera became positive.
The original anti-HMFG monoclonal antibodies, HMFG-1 and HMFG-2 bind to the NPGP complex of the HMFG. These antibodies detected the corresponding antigens in the sera of 30% and 53% of advanced cancer disease cases, respectively. These percentages are low, possibly as a result of the configuration of the assay. In addition, antigenic components with varying molecular weights between 280 and 320 Kdaltons were detected by means of Western blotting in a few of all positive sera detected by the assay. This may indicate either that fragments of the native antigen were found or that the different molecular weight components represent different polymorphic molecules of the antigen. No positives were detected by immunoblotting in any of the normal sera although threshold values were detected by immunoassay.
The DF3 monoclonal antibody also binds to the NPGP complex of the HMFG antigen system. Using this monoclonal antibody a comparison was made between the RIA and ELISA procedures. All yielded increased levels of antigen over the cut off line in more than 70% of the patients with disseminated breast cancer. In contrast thereto, slightly over 5% of normal women had values above the cut off value. Further, 47%, 40% and 27% of patients with ovarian carcinoma pancreatic carcinoma and melanoma were found to have elevated values above the cut-off line whereas ten out of 66 patients with benign liver disease also had elevated values. Patients with visceral breast cancer were found to have a higher frequency of elevated values than those with local or skin recurrences. Results from Western blotting studies were similar to those outlined above for HMFG-1 and HMFG-2 in that a few but not all patients with high immunoassay serum values of the antigen had positive immunoblots.
Another monoclonal antibody, 115D8, raised also against the NPGP complex, was utilized in a sandwich serum assay using the same monoclonal antibody for both layers. The results obtained were similar to the above. About 5% of the samples from normal breasts and benign breast disease showed values above the cut-off line. Breast cancer patients were found positive in 24% of Stage I cases, and in 21%, 43%, and 79% of Stages II through IV cases. 78% positives were found in benign liver disease, kidney disease and in pregnancy cases. A high percentage of ovarian, colorectal, prostate, lung carcinomas, melanoma and lymphoma cases also had elevated values. A good correlation of the marker with the progression or regression of the disease was found in 93% of the cases.
Another early attempt to measure the NPGP complex in circulation was attempted with the F36/22 monoclonal antibody. A similar percentage of positives which were obtained, increases with the severity of disease, as also reported in other assays.
A series of monoclonals were created by another laboratory. Out of a latter group, two named W1 and W9 were selected for also binding to the NPGP complex. Elevated levels of the complex were found in 47% of breast cancer patients with visceral metastases, these being favored over localized metastases. This assay was also positive in 4% of normal cases. Other carcinomas such as colorectal, lung, ovarian, and prostate carcinomas show elevated values of the complex in 12 to 60% of the cases. Recently, another assay using a monoclonal antibody, AB13, detected the NPGP complex in approximately half of advanced breast cancer patients.
The monoclonal antibody DF3 has been used in a commercially available double determinant assay called CA 15-3. Levels above the cut-off line in approximately 80% of advanced cancer patients and in only approximately 30% of primary breast cancer cases were found by this assay. Other authors, in contrast, reported only 13% of primary breast tumors and 72% of disseminated tumors to be positive. In a more detailed study only 24% of breast cancer patients were found to have elevated CA 15-3, while 70% to be positive were found in patients with the disseminated disease. A comparison of CEA and CA 15-3 in both primary and disseminated breast cancer samples, proved the latter to be more sensitive. This commercial assay and the ones discussed above against the heavy molecular weight component of the HMFG or NPGP complexes attain percentages of positives (sensitivities) which are at best similar to those originally reported when the presence of the HME-antigens in the circulation of breast cancer patients was established using polyconal antibodies. The specificities of the assays for the tissue of origin of the tumor are very low (the antigen(s) is almost pan-epithelial), and their specificities for disease conditions are hampered by their high values in hepatic and kidney disease, pregnancy, polymorphic expression of the antigen(s), and the like.
A common feature among the above immunoassays, either utilizing enzymes or radioactivity, is that they indicate a positive correlation between increasing tumor load and higher serum antigen levels. However, a further increase in sensitivity to improve early stage detection of the disease is still necessary. In this regard, an assay using the NPGP complex as a marker and employing the 3E1.2 monoclonal antibody was claimed to have higher sensitivity.
Of a limited number of breast cancer patients up to 68% of early stage patients were found to be positive by this assay whereas only 3% were detected by CA 15-3. The 3E1.2 assay resulted in 18% of positives for benign breast disease resulting then in a low specificity.
Ideally monoclonal antibodies for use in immunoassays would be created against BrE antigen epitopes expressed in breast neoplasias. Alterations in glycoprotein antigens on breast tumor cell membranes have been shown to involve changes in their glycosylation patterns such as substitutions or elongation of oligosaccharide chains without modification of the core sequences. One such monoclonal antibody which is carcinoma specific is B72-3. Its binding to NPGP in benign breast disease tissues was, however, shown later and some normal breast tissue. All the above monoclonal antibody immunoassays for breast cancer rely on serum levels of the NPGP complex. Other antigens, however, have been explored such as the GP-15, Mc3, and Mc8 antigens. The former is a small molecular weight BrE antigen (15 Kdalton) which is present in the cell membrane. It detects mainly, if not exclusively, breast epithelium that has undergone apocrine metaplasia. It has been found, possibly as a result of its selectivity, in the sera of approximately 40% of breast cancer patients. In addition, a small molecular weight antigen (46 Kdalton) of the HMFG system, already detected by in situ radioiodination in the sera of breast cancer patients was measured by a serum immunoassay using a sandwich configuration with monoclonal antibody Mc8 conjugated to biotin as the probing antibody to be finally detected by .sup.125 I-labeled avidin (Salinas et al, Cancer Research 47:907(1987)).
Levels of this antigen were detected in breast cancer patients but not in normal subjects, ovarian carcinomas, colon carcinomas or osteosarcomas. Levels of the Mc3-Mc8 antigen, however, were inversely related to the tumor load. Small tumor loads were 95% positive whereas high tumor loads were 65% positive. This fact was explained by the presence of immune complexes against the antigen whose titer is increased in the high tumor load group. The presence of higher level immune complexes may accelerate clearance of the antigen from blood.
Thus, although most of the above assays employing monoclonal antibodies detect the NPGP complex of the HMFG system, many of them may bind to different epitopes. A heterogeneity of epitopic expression may create the relatively small differences seen among different assays. The diffusely pan-epithelial nature of the NPGP complex as a marker was thus established as shown by its high circulating levels found in other carcinomas, melanomas and even in leukemia. The levels obtained varied depending on the units used in different immunoassays. The percent of positives found at different stages of breast cancer are similar to those originally reported with an assay using polyclonal antibodies to other components of the HMFG. An important drawback of the assays based on the detection of the NPGP complex using monoclonal antibodies is that they lack the specificity of polyclonal assays.
A comparison of the specificities of the CEA assay, an assay detecting the NPGP complex using the Mc5 monoclonal antibody in an antigen displacement, and the original polyclonal antibody assay against HME antigens was made. The polyclonal antibody assay showed very high sensitivity and specificity. It yielded negative values for colon, ovarian, pancreatic, laryngeal and endometrial carcinomas, lymphomas, myelomas, melanomas, and leukemias. Only one case of lung carcinoma showed an elevated value. All normal serum controls were negative, thus showing this assay to have high specificity. Positive serum values for both the NPGP complex and CEA assays were not restricted to breast tumor patient's sera.
Another study reported a higher sensitivity for the polyclonal assay of the HME antigens when compared with the monoclonal antibody assay for the NPGP complex and the CEA test. These three assays were compared in terms of their follow-up ability. The response of one polyclonal antibody assay for HME antigens (cut-off 100 .mu.g/ml) to breast cancer relapse and tumor mass change was quantitated and showed a very sensitive response, far above that demonstrated for the Mc5 assay (cut-off 10 .mu.g/ml) for the NPGP complex. In contrast, the CEA assay either responds slowly or not at all. In clinical cases measurable shrinkage of breast tumor mass was obtained and there was a fast decrease of HME antigens corresponding to a decreased tumor mass brought about by irradiation. The levels of the NPGP complexes remained high and the CEA was unresponsive. In summary, the polyclonal assay for HME antigens has a faster response to changes, and is more accurate in predicting objective changes in tumor mass than the other two assays (CEA and NPGP complex).
Previous studies showed the prognostic power for BrE antigens to be up to 90%. A comparison of the ability of these three assays, the CEA, the NPGP complex, and the HME antigen assays, for predicting relapse was performed by comparing the ability to detect relapse within at least two months in breast cancer patients with no evidence of disease (NED) after an increase of 50% in the serum marker base line by the three methods. The HME antigen method showed a predictability of 73% while the CEA and NPGP complex methods had a 46% predictability. Clearly, among the three, the HME antigen method is the one of choice to establish prognosis due to its high predictive ability and its ability to detect early changes in tumor mass.
U.S. Pat. No. 4,376,110 to David discloses a competitive assay utilizing two antibodies, one of which may be bound to a solid support. The two antibodies compete for the analyte in the sample and may bind to different epitopes of the analyte. The David assay could not be adapted to utilize a fusion protein, even if considered desirable, since the only antigenic substrate present is the analyte, and it cannot be replaced. The prior art assay may be an ELISA or sandwich assay, both of which are depicted in FIGS. 2 and 3 of this patent. The assay of U.S. Pat. No. '110, in addition, requires that the analyte be bound to the free antibody first and, that this free antibody be labeled. Thus the prior art assay produces a soluble labeled analyte-antibody complex, which is totally absent from the present assay, and the amount of label bound to the solid support is proportional to the amount of analyte in the sample whereas in the present assay it is inversely proportional to it.
Other assays utilizing fusion proteins are known in the art. However, they are all different from the present in vitro competitive heterogeneous assay. The following are examples known to the inventors.
Peterhans et al disclose a competitive assay utilizing a fusion protein of .beta.-galactosidase and interferon (Peterhans et al, Analytical Biochemistry 163; 470-475(1987)). That assay uses only one type of antibody, which in this case is bound to a solid support, and a fusion protein and the analyte in the sample, both in solution compete for that antibody. In the Peterhans assay, an anti-interferon-.alpha. monoclonal antibody is attached to the solid support and then incubated with the fusion protein diluted in the sample containing interferon (the analyte). The solid supported material remaining after this step is then incubated in the presence of o-NO.sub.2 -phenyl-glactopyranose (a substrate for .beta.-galactosidase) to determine the amount of fusion protein bound to the antibody and compared with the results obtained by conducting the in the absence of the sample. Although being a competitive assay relying on the use of a fusion protein made of two polypeptides (.beta.-galactosidase and interferon), the prior art assay only utilizes one antibody selectively binding to only one of the two portions of the fusion protein, in this case in solution together with the analyte. FIG. 1 of this patent shows the assay graphically. The prior art assay does not add the fusion protein to a solid phase prior to conducting the competitive portion of the assay is done in the present assay. Both the fusion protein and the sample containing the analyte are added in the same step, and the competitive step is conducted with a solid supported antibody and a fusion protein free in solution whereas the present assay requires a solid supported fusion protein and a free antibody.
In another case, U.S. Pat. No. 4,745,055 to Schenk et al, human surfactant azoprotein (HSA) was determined using an assay configuration similar to the above and a second assay anti-HSA antibody interfered with the .beta.-galactosidase activity of the fusion protein.
Handl et al disclose another competitive assay relying on the utilization of a fusion protein (Handl et al, J.Clin.Microbiol.26:1555-1560(1988)). The fusion protein in this case is composed of .beta.-galactosidase and enterotoxin II and is bound to a solid support. A test sample containing enterotoxin II (the analyte) is added to the solid supported fusion protein, and anti-enterotoxin polyclonal antibodies are then added and both the fusion protein and the enterotoxin II in the sample are allowed to compete for the polyclonal antibodies. The amount of anti-enterotoxin antibody bound to the solid supported fusion protein is determined by adding anti-antibody immunoglobulin which is labeled with alkaline phosphatase. A substrate for the enzyme alkaline phosphatase is then added to the solid supported material and the amount of conversion obtained is compared with that obtained from a similar test conducted in the absence of the test sample. Although this is also a competitive test utilizing a fusion protein and a sample containing the analyte, both capable of binding the antibody, it is different from the assay of the invention in that the fusion protein is bound to the solid support by the .beta.-galactosidase portion thereof, and it only utilizes one type of antibody binding the antigenic peptide portion of the fusion protein. Accordingly, neither the Peterhans nor the Handl assays have the fusion protein in a sandwich form as is the competitive assay of the invention.
Antibodies against HTLV-III were measured in U.S. Pat. No. 4,774,175 to Chang, T. W. et al., using a fusion protein carrying determinants of the virus fixed onto a solid phase.
The above studies show that determination of markers in a clinical setting can be improved by providing an assay of higher sensitivity and/or specificity (above 95%). Such assay will increase clinical diagnosis accuracy by using markers specific for, e.g., different types of cancer and the endocrine system, among others, provided that the markers can be obtained with substantial purity to develop antibodies against them of high avidity and specificity.